Commit 18e4961e authored by Jeremy BLEYER's avatar Jeremy BLEYER

Correct 2018 bugs in modal analysis and hmogenization

parent 874cba7d
......@@ -106,15 +106,13 @@ can be solved using the ``SLEPcEigenSolver``. ::
eigensolver = SLEPcEigenSolver(K, M)
eigensolver.parameters['problem_type'] = 'gen_hermitian'
eigensolver.parameters["spectrum"] = "smallest real"
eigensolver.parameters['spectral_transform'] = 'shift-and-invert'
eigensolver.parameters['spectral_shift'] = 0.
The problem type is specified to be a generalized eigenvalue problem with
Hermitian matrices. By default, SLEPc computes the largest eigenvalues, here
we instead look for the smallest eigenvalues (they should all be real). To
improve convergence of the eigensolver for finding the smallest eigenvalues
(by default it computes the largest ones), a spectral transform is performed
Hermitian matrices. By default, SLEPc computes the largest eigenvalues. Here
we instead look for the smallest eigenvalues (they should all be real). A
spectral transform is therefore performed
using the keyword ``shift-invert`` i.e. the original problem is transformed into
an equivalent problem with eigenvalues given by :math:`\dfrac{1}{\lambda - \sigma}`
instead of :math:`\lambda` where :math:`\sigma` is the value of the spectral shift.
......
This diff is collapsed.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment